“The Fiduciary Duty Dilemma: Exploring the Legality of AI-Assisted Decision Making by Directors”
The term ‘Artificial Intelligence’ (“AI”) was first coined by an American computer scientist John McCarthy in 1956 where he defined it as, “the science and engineering of making intelligent machines, especially intelligent computer.
”[1] After its promulgation and subsequent adoption, AI has galloped and has in turn created opportunities and challenges for companies, organizations, governments all over the world. It should be critically understood that the advancement of AI has presented regulators and governments across the globe with multifarious challenges. One such challenge faced by regulators nowadays is the use of AI by directors to make reasoned and informed decisions to ensure and maintain the effectiveness and efficiency of their organizations and companies.
In 2014, a venture capital firm – Deep Knowledge Ventures, appointed VITAL[2], a machine learning program, to its board as a member with observer status.[3] This revolutionised the concept of AI and spoke volumes regarding its use and dependency by companies. However, VITAL did not qualify as a ‘natural person’ under the legal requirement of corporate directors in Hong Kong.[4] This inspired the debate whether AI can be appointed as a director or hold office in a company. Subsequently, Tieto, a Finnish IT company, appointed Alicia T, an AI application, on its leadership team with voting rights to lead the new data driven business unit in 2016.[5] However, it is very well placed under the UK[6], US[7], and Indian[8] company law that only natural persons or individuals can be appointed as directors in companies.
Fiduciary duties of Directors
It is well noted that each member of the board of directors of a company have fiduciary duties and responsibilities and any decision made by the board of directors directly impacts their company and its stakeholders, which include shareholders, employees, customers, and the public at large (for listed entities). These duties include the duty of care, duty of loyalty, duty to act in good faith, and duty to disclose all relevant information in certain specific form and manner depending upon the jurisdiction and applicable laws. If the board of directors use AI tools for arriving at a business decision, whether such reliance on the AI tools would mean that the directors are in breach of their duties prescribed under the applicable laws? To answer this query, it is pertinent to analyse in what circumstances and to what extent directors should rely on the outputs provided by AI tools.
The need of AI tools – ‘Choice Overload’ and ‘Analysis Paralysis’
To understand directors’ fiduciary liability and responsibility while using AI to make decisions, it is essential to understand the nature of AI and how it works. AI is a system that can learn and improve over time, based on data inputs and algorithms. It can be used to analyse large amounts of data, identify patterns and trends, and make predictions or recommendations. The use of AI tools has increased substantially in Directors of companies nowadays use AI to assist them in their day-to-day functions to mitigate agency problems (refers to a corporate governance situation where the board of directors have to prioritize the interest of the company over the interest of the shareholders of the company and therefore, the decision-making process becomes extremely critical). Such agency problems also create a situation of ‘choice overload’[9] and ‘analysis paralysis’[10] for the directors where they use AI tools for deriving the relevant data to justify their decisions. In such circumstances, delegation to AI tools can prove helpful to solve problems where decisions are to be taken on the basis of a large quantum of data and information due to which directors are unable to take swift and efficient decisions which, otherwise is a time consuming and tough exercise. However, how AI can present opportunities and predicaments to company directors in the legal context still needs to be analysed and understood.
Legal Framework
While there is no express law in India to regulate the use of AI by directors, the Companies Act, 2013 (“CA”) expressly sets out certain duties[11] that directors that must adhere to while discharging their duties and responsibilities. However, section 166(3) of the CA provides that, “A director of a company shall exercise his duties with due and reasonable care, skill and diligence and shall exercise independent judgement.”[12] It can be reasonably inferred that every director shall exercise independent judgement while discharging their duties and responsibilities. The Indian courts have provided very limited dicta[13] as to what constitutes independent judgement, whereas the UK courts have gone a step ahead and postulated that independent judgment means ‘the judgement of the director is independent in the sense of it being his own judgement’[14]. Therefore, it can be said that directors can use AI tools for deriving the relevant data as well as information to arrive at an independent business decision, as long as the decision-making is not solely delegated to such AI tools.
Doctrine of ‘Business Judgement Rule’
While other jurisdictions have only started to regulate the use of AI, the evolution and advancement of AI has not stopped. In 2017, the European parliament has emphasized the need to create a robust legal framework to oversee the formation and evolution of AI.[15] The US legal system has taken a step ahead by instilling the business judgement rule which is a basic liability rule that safeguards the board of directors against frivolous legal allegations as to the way it conducts business. Directors shall not incur any liability towards their decisions if it was an informed decision; there was no conflict of interest; and lastly if the directors acted in good faith.[16] Although, Indian law does not have a express ‘business judgement rule’, the CA provides exemptions to directors from penalties in cases of negligence, default, breach of duty, misfeasance or breach of trust, if the director can prove he acted honestly and reasonably, and had regard towards all the circumstances presented.[17] It is also important to note that this doctrine has been accepted by Indian courts for example, the Supreme Court in the case of Miheer Mafatlal, held that courts shall not interfere if the conduct was “just, fair and reasonable from the point of prudent men of business taking a commercial decision beneficial to the class represented by them.”[18] The decision of the apex court was indeed a welcome move towards adoption of the business judgement rule. It can be inferred that in case the board of directors use AI tools for limited purpose of data collection to support their judgment, such decisions should also be included under this ‘business judgment rule’.
Conclusion
Director’s responsibilities inherently demand the exercise of judgment. Corporate governance comprises of both legal and social norms, however, it may be difficult to frame an express legal rules which can address every situation that a director or the board collectively may find itself in. If directors use AI to take decisions, they should ensure the reliability and accuracy of AI systems to act in the best interests of the company while being transparent about the use of AI and its potential impact. As AI continues to become more prevalent in business, it is essential that directors understand their responsibilities and take steps to ensure that the decisions taken through AI do not negatively impact the progress of the company and is in consonance with the applicable rules and regulations.
AUTHORS:
Ambuj Sonal, Associate Partner, Dentons Link Legal
Tanay Jha, Associate, Dentons Link Legal
Footnotes
[1] https://www-formal.stanford.edu/jmc/whatisai.pdf
[2] Validating Investment Tool for Advancing Life Science
[3] https://www.globenewswire.com/news-release/2014/05/13/635881/10081467/en/Deep-Knowledge-Venture-s-Appoints-Intelligent-Investment-Analysis-Software-VITAL-as-Board-Member.html and https://asia.nikkei.com/Business/Artificial-intelligence-gets-a-seat-in-the-boardroom
[4] Clause 457 of the Hong Kong Companies Ordinance 2014, https://www.elegislation.gov.hk/hk/cap622
[5] https://www.bloomberg.com/press-releases/2016-10-17/tieto-the-first-nordic-company-to-appoint-artificial-intelligence-to-the-leadership-team-of-the-new-data-driven-businesses-unit
[6] Section 155(1) of the UK Companies Act, 2006
[7] Delaware General Corporation Law, § 141 (b), https://delcode.delaware.gov/title8/c001/sc04/
[8] Section 152 of the Companies Act, 2013 and Gautam Mehra v. Union of India & Ors. LNIND 2020 CAL 3489
[9] ‘Choice Overload’ refers to the paradigm of extensive choices available to an individual, which may prove to be desirable initially but may prove to be demotivating in the end.
[10] ‘Analysis Paralysis’ refers to the situation where overanalysing data and information may lead to the inability to take or proceed with a decision.
[11] Section 166 of the Companies Act, 2013
[12] Section 166(3) of the Companies Act, 2013
[13] Tata Consultancy Services Ltd. v. Cyrus Investments Pvt. Ltd. and Ors., (2021) 9 SCC 449
[14] Lord Goldsmith, Lords Grand Committee, 6 February 2006, column 282, Hansard.
[15] https://www.theguardian.com/technology/2017/jan/12/give-robots-personhood-status-eu-committee-argues
[16] Aronson v. Lewis, 473 A. 2d 805
[17] Section 463(1) of the Companies Act, 2013
[18] Miheer H. Mafatlal v. Mafatlal Industries Ltd., JT 1996 (8) 205